• No results found

Appendix B: Additional Figures and Tables Not for Publication

In document Edinburgh Research Explorer (Page 46-58)

Table 2A: Tests for Differences Between Subgroups

Column Column Column [2] = [3] [4] = [5] [6] = [7]

Competition (=1) * Week 2-5 0.577 0.686 0.611

Competition (=1) * Week 2-5 (wild-p) 0.595 0.681 0.687

Competition (=1) * Week 6 0.164 0.985 0.216

Competition (=1) * Week 6 (wild-p) 0.186 1.019 0.240 Individual Incentive (=1) * Week 2-5 0.543 0.316 0.020 Individual Incentive (=1) * Week 2-5 (wild-p) 0.571 0.316 0.076 Individual Incentive (=1) * Week 6 0.871 0.269 0.067 Individual Incentive (=1) * Week 6 (wild-p) 0.893 0.322 0.132

First Group in Column Heading Female FSM Year 2

Second Group in Column Heading Male Non-FSM Year 5

The table contains the p-values for the tests of whether the coefficient on the variables in Table 2 for the two columns listed are equal. Wild p-values shown are estimated following Cameron, Gelbrach, Miller (2008) using 1000 replications. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6.

Table 3A: Tests for Differences Between Subgroups

Column Column Column [2] = [3] [4] = [5] [6] = [7]

Competition (=1) * Week 2-5 0.240 0.456 0.972

Competition (=1) * Week 2-5 (wild-p) 0.276 0.573 0.911

Competition (=1) * Week 6 0.473 0.951 0.570

Competition (=1) * Week 6 (wild-p) 0.529 0.907 0.637 Individual Incentive (=1) * Week 2-5 0.072 0.729 0.000 Individual Incentive (=1) * Week 2-5 (wild-p) 0.154 0.733 0.002 Individual Incentive (=1) * Week 6 0.205 0.444 0.000 Individual Incentive (=1) * Week 6 (wild-p) 0.252 0.611 0.002

First Group in Column Heading Female FSM Year 2

Second Group in Column Heading Male Non-FSM Year 5

The table contains the p-values for the tests of whether the coefficient on the variables in Table 3 for the two columns listed are equal. Wild p-values shown are estimated following Cameron, Gelbrach, Miller (2008) using 1000 replications. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6.

Table 4A: Tests for Differences Between Subgroups

Column Column Column [2] = [3] [4] = [5] [6] = [7]

Competition (=1) * Week 2-5 0.324 0.204 0.831

Competition (=1) * Week 2-5 (wild-p) 0.376 0.284 0.847

Competition (=1) * Week 6 0.229 0.202 0.299

Competition (=1) * Week 6 (wild-p) 0.248 0.316 0.338 Individual Incentive (=1) * Week 2-5 0.745 0.437 0.001 Individual Incentive (=1) * Week 2-5 (wild-p) 0.775 0.452 0.020 Individual Incentive (=1) * Week 6 0.965 0.364 0.012 Individual Incentive (=1) * Week 6 (wild-p) 0.969 0.378 0.068

First Group in Column Heading Female FSM Year 2

Second Group in Column Heading Male Non-FSM Year 5

The table contains the p-values for the tests of whether the coefficient on the variables in Table 4 for the two columns listed are equal. Wild p-values shown are estimated following Cameron, Gelbrach, Miller (2008) using 1000 replications. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6.

Table 5A: Tests for Differences Between Subgroups

Column Column Column [2] = [3] [4] = [5] [6] = [7]

Competition (=1) * Week 2-5 0.362 0.444 0.608

Competition (=1) * Week 2-5 (wild-p) 0.360 0.468 0.679

Competition (=1) * Week 6 0.528 0.441 0.292

Competition (=1) * Week 6 (wild-p) 0.601 0.513 0.324 Individual Incentive (=1) * Week 2-5 0.139 0.768 0.000 Individual Incentive (=1) * Week 2-5 (wild-p) 0.280 0.765 0.000 Individual Incentive (=1) * Week 6 0.322 0.727 0.000 Individual Incentive (=1) * Week 6 (wild-p) 0.362 0.695 0.000

First Group in Column Heading Female FSM Year 2

Second Group in Column Heading Male Non-FSM Year 5

The table contains the p-values for the tests of whether the coefficient on the variables in Table 5 for the two columns listed are equal. Wild p-values shown are estimated following Cameron, Gelbrach, Miller (2008) using 1000 replications. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6.

Table 7A: Tests for Differences Between Subgroups

Column Column Column [2] = [3] [4] = [5] [6] = [7]

Competition (=1) * 6 months later 0.152 0.223 0.490

Competition (=1) * 6 months later (wild-p) 0.206 0.282 0.484 Individual Incentive (=1) * 6 months later 0.601 0.406 0.332 Individual Incentive (=1) * 6 months later (wild-p) 0.587 0.478 0.448

First Group in Column Heading Female FSM Year 2

Second Group in Column Heading Male Non-FSM Year 5

The table contains the p-values for the tests of whether the coefficient on the variables in Table 7 for the two columns listed are equal. Wild p-values shown are estimated following Cameron, Gelbrach, Miller (2008).

The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6 or 6 months later.

Table 8A: Tests for Differences Between Subgroups

Column Column Column [2] = [3] [4] = [5] [6] = [7]

Competition (=1) * 6 months later 0.581 0.044 0.865

Competition (=1) * 6 months later (wild-p) 0.631 0.144 0.887 Individual Incentive (=1) * 6 months later 0.940 0.843 0.053 Individual Incentive (=1) * 6 months later (wild-p) 0.927 0.859 0.112

First Group in Column Heading Female FSM Year 2

Second Group in Column Heading Male Non-FSM Year 5

The table contains the p-values for the tests of whether the coefficent on the variables in Table 8 for the two columns listed are equal. Wild p-values shown are estimated following Cameron, Gelbrach, Miller (2008) using 1000 replications. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6 or 6 months later.

Figure A2: Example food knowledge test

Figure B1: Proportion of pupils eating more than half a fruit or vegetable

a) Full Sample

a) Sample with less than 100% Choice in Baseline

Table B1: Descriptive Statistics of Local Education Authorities

(1) (2) (3) (4)

Participate Collaborate but Did not p-value not participate collaborate

Contacted on Friday 0.75 0.41 0.47 0.163

Contacted by J James 0.58 0.53 0.50 0.795

Household Income/100 6.10 7.23 6.72 0.138

% FSM 0.15 0.21 0.17 0.164

Number of Schools/100 2.27 1.50 1.35 0.037

% 5 Fruit & Veg a day 24.5 25.9 26.0 0.603

% Overweight & Obese reception 23.6 23.2 23.0 0.714

% Obese reception 9.94 10.4 9.74 0.330

% Overweight & Obese yr6 33.1 34.4 33.0 0.309

% Obese yr6 18.5 20.1 18.7 0.180

Smoking 25.6 24.7 24.6 0.794

Binge Drinking 20.2 17.1 18.1 0.195

Key stage 1: Avg point score 0.15 0.15 0.15 0.879

Key stage 2: Avg point score 0.28 0.28 0.28 0.894

Per pupil spending 2010/11 4307 4806 4486 0.109

% change in per pupil spending 2010/11 0.04 0.04 0.04 0.778

% LA spending change 2010/11 -0.05 -0.05 -0.05 0.689

Female CEO of the council 0.25 0.31 0.25 0.859

Female Director of Children Services 0.67 0.41 0.51 0.405

Female Leader of Healthy Schools 0.75 0.88 0.82 0.659

% of Labour Councillors 0.32 0.39 0.33 0.650

% of Conservative Councillors 0.43 0.43 0.42 0.983

Labour controlled council 0.25 0.29 0.29 0.958

Conservative controlled council 0.42 0.47 0.41 0.878

Ofsted Score 2.34 2.31 2.28 0.405

Ofsted Health Score 1.69 1.72 1.70 0.707

Catering pp/100 0.80 0.83 0.82 0.989

Energy costs pp/100 0.69 0.68 0.65 0.610

Total school Income pp/1000 4.30 4.52 4.32 0.674

Teaching costs pp/1000 2.10 2.18 2.13 0.756

P-values in column 4 come from a test of equality of the 3 group means between local authorities that participated, those that collaborated (by providing names of schools), and those that did not collaborate.

Local authorities were randomly contacted on two days on Friday 2nd July and Monday 5th July and by J. James or M. Belot. Income is the average weekly total household income (£) divided by 100, FSM is the percentage of children who are eligible for free school meals. % Eat 5 Fruit & Veg a day is the proportion of adults defined to be consumers of 5 or more fruit and vegetables if they had reported that they had consumed 5 or more portions of fruit and vegetables on the previous day. Binge drinking is the proportion of adult binge drinkers defined if they reported that in the last week they had drunk 8 or more units of alcohol if they were a man, or 6 or more units of alcohol if they were a woman, on any one day or more. Smoking is the proportion of individuals in a local authority who reported that they were a ‘current cigarette smoker’ in the Health Survey for England. Overweight and Obese reception is the percentage of pupils in the local authority who were overweight or obese when they entered primary school aged 4 or 5.

Year 6 is the final year of school when the pupils are aged 10 or 11. The average point score (APS) of the key stage 1 test and key stage 2 point score are for tests taken in primary school. The points are awarded per subject per pupil along the following lines: working below the level of the test or not awarded 15, level 2 receives 15 points, level 3 gets 21, level 4 gets 27 and 33 points is allocated for level 5. The APS is then calculated using the following: (Total points for English + Total points for maths + Total points for science) / (Total number of eligible pupils for each subject). This is then rescaled by dividing by 100. Per pupil spending in 2010/11, the yearly increase in per pupil spending, and the overall change in the spending of the local authority. Labour Party and Conservative Party councillors on the council defined at the most recent election since July 2010. Ofsted is (the government school inspector) average score of the schools in the local authority. Schools are inspected and judged on the following question: “How effective, efficient and inclusive is the provision of education, integrated care and any extended services in meeting the needs of learners?” With ratings given of: 1. Outstanding 2. Good 3. Satisfactory 4. Inadequate. Ofsted Health Score is based on the following question: “Learners are encouraged and enabled to eat and drink healthily” using the same 1 to 4 scale. Average catering costs (including staff costs), energy, teaching and total school income are per pupil averages at the local authority level and are rescaled as indicated.

50

Table B2 Comparison of participating schools from the pool of selected schools

Total School Income per pupil/1000 4.17 4.16 0.641

Absenteeism (% on census day) 0.05 0.05 0.682

Catering costs per pupil/100 0.96 0.73 0.303

% English and Maths above level 4 KS2 0.76 0.76 0.949 Average point score Maths and English 0.28 0.28 0.396

Food for life 0.31 0.21 0.501

Ofsted Score 2.09 2.29 0.521

Ofsted health Score 1.53 1.43 0.604

FSM Medium 0.29 0.29 0.975

FSM Low 0.65 0.64 0.988

Teaching Costs per pupil/1000 2.05 2.17 0.246

Energy costs per pupil/100 0.64 0.87 0.961

Competition treatment 0.29 0.43 0.368

Individual treatment 0.32 0.36 0.822

Control 0.39 0.21 0.260

Schools 31 15

Columns 1 and 2 show the mean values at the school level. Column 3 is the p-value of (Prob>z, where z is the test statistic) from an Mann-Whitney U test. Ofsted is (the government school inspector) average score of the schools in the local authority.

Schools are inspected and judged on the following question: “How effective, efficient and inclusive is the provision of education, integrated care and any extended services in meeting the needs of learners?” With ratings given of: 1. Outstanding 2. Good 3. Satisfactory 4. Inadequate. Ofsted Health Score is based on the following question: “Learners are encouraged and enabled to eat and drink healthily” using the same 1 to 4 scale. Average catering costs (including staff costs), energy, teaching and total school income are per pupil averages at the local authority level and are rescaled as indicated. FSM Band - The three broad bands used to group pupils eligible for FSM are: Low: less than 20%, Medium: 20.01-35% and High: greater than 35% (omitted catergory). Columns (1)-(3) present estimates using whether a school was selected by the LEA. Column (3) excludes “Avg Eng/Math Score” but uses the same sample in column (2). Column (4) and (5) use whether a school started and completed the experimental intervention.

Table B3: Effect on Eating More than Half for Overall Sample and Its Subgroups

Dependent Variable (=1) if Student Ate More than Half a Healthy Item

All Females Males FSM Non-FSM Year 2 Year 5

[1] [2] [3] [4] [5] [6] [7]

Competition (=1) * Week 2-5 0.114* 0.129 0.096 0.107 0.120 0.096 0.133**

(0.063) (0.084) (0.079) (0.086) (0.072) (0.108) (0.063) [0.194] [0.178] [0.288] [0.272] [0.144] [0.438] [0.070]

Competition (=1) * Week 6 0.082 0.099 0.061 0.124 0.078 0.108 0.062 (0.073) (0.104) (0.073) (0.086) (0.088) (0.111) (0.083) [0.354] [0.416] [0.490] [0.168] [0.420] [0.418] [0.505]

Individual Incentive (=1) * Week 2-5 0.054 0.051 0.053 0.008 0.057 -0.054 0.219***

(0.060) (0.076) (0.067) (0.072) (0.066) (0.072) (0.048) [0.464] [0.561] [0.438] [0.927] [0.452] [0.498] [0.014]

Individual Incentive (=1) * Week 6 0.008 0.040 -0.023 -0.010 0.005 -0.068 0.143 (0.075) (0.091) (0.078) (0.101) (0.083) (0.083) (0.090) [0.893] [0.695] [0.813] [0.915] [0.989] [0.488] [0.172]

Constant 0.599*** 0.628*** 0.567*** 0.592*** 0.606*** 0.602*** 0.588***

(0.022) (0.029) (0.026) (0.029) (0.025) (0.032) (0.021)

P-Value: Comp = Ind Incentive Week 2-5 0.410 0.356 0.638 0.320 0.437 0.109 0.193 P-Value: Comp = Ind Incentive Week 2-5 (wild) 0.488 0.428 0.687 0.360 0.460 0.164 0.256 P-Value: Comp = Ind Incentive Week 6 0.327 0.502 0.340 0.212 0.387 0.049 0.294 P-Value: Comp = Ind Incentive Week 6 (wild) 0.446 0.607 0.390 0.256 0.444 0.054 0.352 Observations 14,714 7,719 6,994 2,495 11,838 7,916 6,798

R-squared 0.012 0.012 0.013 0.010 0.011 0.009 0.025

Number of pupils 638 328 310 114 509 343 295

Robust standard errors clustered at school level are included in parentheses; * sig at 10%, ** sig at 5%, *** sig at 1%. Wild P-Values are shown in brackets and are estimated following Cameron, Gelbach, Miller (2008). “Eat more than half ” is a dummy variable equal to one if the pupil eats at least half of a portion of fruits or vegetables on that day. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6. Column [2] includes only female and column [3] contains only males. Column [4] includes only pupils who are eligible for free school meals (FSM) and column [5] contains those

52

Table B4: Effect on Eating More Than Half for Sample with Week 1 less than 100% Choice and Its Subgroups Dependent Variable (=1) if Student Ate Mopre than Half a Healthy Item

All Females Males FSM Non-FSM Year 2 Year 5

[1] [2] [3] [4] [5] [6] [7]

Competition (=1) * Week 2-5 0.190** 0.145 0.218** 0.268** 0.175** 0.141 0.203**

(0.076) (0.095) (0.088) (0.114) (0.076) (0.100) (0.087) [0.024] [0.178] [0.042] [0.104] [0.038] [0.230] [0.036]

Competition (=1) * Week 6 0.117* 0.074 0.143** 0.245** 0.086 0.119 0.094 (0.066) (0.102) (0.064) (0.095) (0.068) (0.069) (0.099)

[0.126] [0.501] [0.052] [0.058] [0.288] [0.172] [0.404]

Individual Incentive (=1) * Week 2-5 0.078 0.001 0.130 0.096 0.061 -0.193*** 0.216***

(0.068) (0.091) (0.082) (0.171) (0.069) (0.063) (0.063) [0.318] [0.973] [0.292] [0.695] [0.466] [0.016] [0.008]

Individual Incentive (=1) * Week 6 -0.006 -0.024 0.003 0.049 -0.030 -0.326*** 0.133 (0.096) (0.102) (0.118) (0.272) (0.097) (0.073) (0.106)

[0.979] [0.795] [0.979] [0.617] [0.773] [0.004] [0.270]

Constant 0.342*** 0.372*** 0.314*** 0.231*** 0.363*** 0.291*** 0.381***

(0.025) (0.030) (0.029) (0.047) (0.025) (0.031) (0.027)

P-value for Competition=Individual for Wks 2-5 0.199 0.104 0.420 0.391 0.183 0.001 0.883 P-value for Competition=Individual for Wks 2-5 (wild) 0.220 0.134 0.513 0.511 0.228 0.008 0.879 P-value for Competition=Individual for Wk 6 0.166 0.121 0.274 0.507 0.156 0.000 0.692 P-value for Competition=Individual for Wk 6 (wild) 0.210 0.110 0.322 0.555 0.124 0.000 0.665

Observations 5,466 2,583 2,883 799 4,476 2,360 3,106

R-squared 0.057 0.065 0.052 0.082 0.051 0.072 0.058

Number of pupils 215 102 113 29 179 93 122

Robust standard errors clustered at school level are included in parentheses; * sig at 10%, ** sig at 5%, *** sig at 1%. Wild P-Values are shown in brackets and are estimated following Cameron, Gelbach, Miller (2008). “Eat more than half ” is a dummy variable equal to one if the pupil eats at least half of a portion of fruits or vegetables on that day. The incentives were in place during weeks 2 to 5 (inclusive) but were not in place during week 6. Column [2] includes only female and

53

Table B5: Effect on Attendance On Overall Sample and Its Subgroups

Dependent Variable (=1) if Student Attended School

All Females Males FSM Non-FSM Year 2 Year 5

[1] [2] [3] [4] [5] [6] [7]

Competition (=1) * Week 2-5 0.017 0.002 0.037* 0.029 0.015 0.021 0.016 (0.014) (0.021) (0.018) (0.051) (0.016) (0.017) (0.021)

[0.276] [0.897] [0.068] [0.621] [0.396] [0.304] [0.559]

Competition (=1) * Week 6 -0.009 -0.023 0.014 -0.014 -0.006 -0.011 -0.004 (0.017) (0.026) (0.020) (0.061) (0.015) (0.027) (0.028)

[0.655] [0.412] [0.474] [0.811] [0.675] [0.645] [0.833]

Individual Incentive (=1) * Week 2-5 0.023 0.009 0.040* 0.002 0.029 0.015 0.032 (0.022) (0.029) (0.023) (0.042) (0.026) (0.018) (0.037)

[0.414] [0.783] [0.116] [0.931] [0.306] [0.444] [0.482]

Individual Incentive (=1) * Week 6 -0.022 -0.031 -0.007 -0.061* -0.007 -0.007 -0.035 (0.048) (0.050) (0.050) (0.032) (0.049) (0.020) (0.099)

[0.733] [0.581] [0.937] [0.104] [0.865] [0.717] [0.809]

Constant 0.945*** 0.945*** 0.946*** 0.971*** 0.938*** 0.956*** 0.934***

(0.007) (0.009) (0.008) (0.014) (0.008) (0.007) (0.013) P-value for Competition=Individual for Wks 2-5 0.800 0.814 0.877 0.411 0.551 0.790 0.634 P-value for Competition=Individual for Wks 2-5 (wild) 0.831 0.859 0.917 0.482 0.579 0.837 0.689 Observations 16,472 8,548 7,917 2,843 13,200 8,596 7,876

R-squared 0.003 0.002 0.004 0.009 0.002 0.001 0.007

Number of pupils 643 331 312 115 513 345 298

Robust standard errors clustered at school level are included in parentheses; * sig at 10%, ** sig at 5%, *** sig at 1%. Wild P-Values are shown in brackets and are estimated following Cameron, Gelbach, Miller (2008). Column [2] includes only female and column [3] contains only males.

Column [4] includes only pupils who are eligible for free school meals (FSM) and column [5] contains those pupils not eligible for FSM; there are 15 pupils for whom we are missing FSM status. Column [6] contains only pupils in Year 2 and column [7] contains only pupils in Year 5.

54

Table B6: Effect on Attendance for Sample with Week 1 less than 100% Choice and Its Subgroups Dependent Variable (=1) if Student Attended School

All Females Males FSM Non-FSM Year 2 Year 5

[1] [2] [3] [4] [5] [6] [7]

Competition (=1) * Week 2-5 -0.015 -0.063 0.030 0.046** -0.025 -0.032 0.011 (0.023) (0.037) (0.027) (0.019) (0.028) (0.038) (0.034)

[0.563] [0.322] [0.380] [0.076] [0.424] [0.424] [0.785]

Competition (=1) * Week 6 -0.062** -0.130*** 0.010 -0.003 -0.067** -0.081* -0.034 (0.022) (0.041) (0.034) (0.036) (0.029) (0.042) (0.036)

[0.034] [0.04] [0.765] [0.777] [0.070] [0.054] [0.394]

Individual Incentive (=1) * Week 2-5 0.062 0.041 0.078** 0.040*** 0.065 0.057 0.063 (0.040) (0.060) (0.035) (0.005) (0.044) (0.070) (0.048)

[0.204] [0.533] [0.066] [0.124] [0.208] [0.440] [0.386]

Individual Incentive (=1) * Week 6 0.045 -0.020 0.091** -0.100 0.059 0.028 0.053 (0.041) (0.071) (0.042) (0.059) (0.044) (0.096) (0.034)

[0.266] [0.823] [0.014] [0.507] [0.206] [0.789] [0.240]

Constant 0.909*** 0.901*** 0.915*** 0.980*** 0.894*** 0.931*** 0.892***

(0.010) (0.014) (0.011) (0.007) (0.012) (0.016) (0.014) P-value for Competition=Individual for Wks 2-5 0.0443 0.0496 0.256 0.757 0.0324 0.163 0.233 P-value for Competition=Individual for Wks 2-5 (wild) 0.130 0.228 0.306 0.785 0.136 0.150 0.430

Observations 6,085 2,870 3,210 838 5,047 2,582 3,503

R-squared 0.008 0.016 0.006 0.014 0.010 0.006 0.011

Number of pupil 220 105 115 30 183 95 125

Robust standard errors clustered at school level are included in parentheses; * sig at 10%, ** sig at 5%, *** sig at 1%. Wild P-Values are shown in brackets and are estimated following Cameron, Gelbach, Miller (2008). Column [2] includes only female and column [3] contains only males.

Column [4] includes only pupils who are eligible for free school meals (FSM) and column [5] contains those pupils not eligible for FSM; there are 15 pupils for whom we are missing FSM status. Column [6] contains only pupils in Year 2 and column [7] contains only pupils in Year 5.

55

Table B7: Effect on Bringing Packed Lunch On Overall Sample and Its Subgroups

Dependent Variable (=1) if Student Brought a Packed Lunch

All Females Males FSM Non-FSM Year 2 Year 5

[1] [2] [3] [4] [5] [6] [7]

Competition (=1) * Week 2-5 0.000 0.000 -0.001 0.014 0.001 0.008 -0.014 (0.021) (0.032) (0.023) (0.039) (0.025) (0.034) (0.025)

[0.993] [0.995] [0.957] [0.737] [0.951] [0.849] [0.635]

Competition (=1) * Week 6 -0.038 -0.065 -0.003 0.008 -0.042 -0.063 -0.020 (0.030) (0.046) (0.033) (0.044) (0.038) (0.043) (0.036)

[0.220] [0.176] [0.923] [0.883] [0.332] [0.202] [0.621]

Individual Incentive (=1) * Week 2-5 -0.013 -0.001 -0.022 -0.038* 0.004 -0.014 -0.014 (0.025) (0.035) (0.020) (0.021) (0.024) (0.033) (0.037)

[0.569] [1.02] [0.394] [0.200] [0.827] [0.681] [0.815]

Individual Incentive (=1) * Week 6 -0.041 -0.037 -0.040 -0.057 -0.021 -0.078* -0.008 (0.036) (0.052) (0.029) (0.042) (0.036) (0.043) (0.055)

[0.256] [0.509] [0.268] [0.258] [0.587] [0.128] [0.919]

Constant 0.499*** 0.489*** 0.511*** 0.187*** 0.566*** 0.461*** 0.539***

(0.008) (0.011) (0.008) (0.013) (0.009) (0.009) (0.012) P-value for Competition=Individual for Wks 2-5 0.525 0.968 0.421 0.255 0.919 0.0684 0.996 P-value for Competition=Individual for Wks 2-5 (wild) 0.583 1.035 0.482 0.306 0.865 0.092 0.957 Observations 14,575 7,622 6,953 2,501 11,671 7,348 7,227

R-squared 0.002 0.002 0.002 0.004 0.002 0.003 0.002

Number of pupils 623 322 301 110 498 329 294

Robust standard errors clustered at school level are included in parentheses; * sig at 10%, ** sig at 5%, *** sig at 1%. Wild P-Values are shown in brackets and are estimated following Cameron, Gelbach, Miller (2008). Column [2] includes only female and column [3] contains only males.

Column [4] includes only pupils who are eligible for free school meals (FSM) and column [5] contains those pupils not eligible for FSM; there are 15 pupils for whom we are missing FSM status. Column [6] contains only pupils in Year 2 and column [7] contains only pupils in Year 5.

56

Table B8: Effect on Bringing Packed Lunch for Sample with Week 1 less than 100% Choice and Its Subgroups Dependent Variable (=1) if Student Brought a Packed Lunch

All Females Males FSM Non-FSM Year 2 Year 5

[1] [2] [3] [4] [5] [6] [7]

Competition (=1) * Week 2-5 0.007 -0.021 0.040 0.033 -0.000 0.020 -0.019 (0.023) (0.040) (0.026) (0.118) (0.023) (0.029) (0.041)

[0.719] [0.641] [0.124] [0.783] [0.991] [0.543] [0.657]

Competition (=1) * Week 6 -0.004 -0.076 0.080** -0.006 0.003 -0.039 0.005 (0.036) (0.071) (0.030) (0.121) (0.043) (0.071) (0.058)

[0.957] [0.348] [0.032] [0.985] [0.971] [0.515] [0.925]

Individual Incentive (=1) * Week 2-5 0.036 0.054* 0.022 0.007 0.053* 0.060 0.027 (0.025) (0.030) (0.038) (0.005) (0.028) (0.054) (0.022)

[0.204] [0.182] [0.643] [0.430] [0.072] [0.595] [0.408]

Individual Incentive (=1) * Week 6 0.018 0.044 -0.003 -0.017 0.048 -0.039 0.050 (0.046) (0.076) (0.041) (0.014) (0.041) (0.072) (0.057)

[0.751] [0.651] [0.941] [0.505] [0.350] [0.527] [0.645]

Constant 0.532*** 0.527*** 0.536*** 0.355*** 0.564*** 0.509*** 0.549***

(0.009) (0.015) (0.011) (0.042) (0.009) (0.013) (0.012) P-value for Competition=Individual for Wks 2-5 0.318 0.0518 0.646 0.825 0.0749 0.466 0.262 P-value for Competition=Individual for Wks 2-5 (wild) 0.384 0.112 0.697 0.821 0.100 0.781 0.302

Observations 5,376 2,555 2,821 771 4,412 2,195 3,181

R-squared 0.001 0.004 0.002 0.002 0.001 0.004 0.002

Number of pupils 214 102 112 29 178 93 121

Robust standard errors clustered at school level are included in parentheses; * sig at 10%, ** sig at 5%, *** sig at 1%. Wild P-Values are shown in brackets and are estimated following Cameron, Gelbach, Miller (2008). Column [2] includes only female and column [3] contains only males.

Column [4] includes only pupils who are eligible for free school meals (FSM) and column [5] contains those pupils not eligible for FSM; there are 15 pupils for whom we are missing FSM status. Column [6] contains only pupils in Year 2 and column [7] contains only pupils in Year 5.

57

In document Edinburgh Research Explorer (Page 46-58)

Related documents